Machine Learning Predicts Pollution-Linked ACS Mortality
Researchers led by Sazzli Kasim publish in Scientific Reports this week a study analyzing 14,145 Malaysian acute coronary syndrome cases from 2006–2017, combining National Cardiovascular Disease Database clinical records with daily air-quality metrics. A random forest model achieved AUC 0.843 versus TIMI’s 0.791 (STEMI) and 0.565 (NSTEMI), with net reclassification improvements of 8.71% (STEMI) and 86.94% (NSTEMI); SHAP flagged NOx and O3 as top predictors, indicating pollution-aware models improve mortality prediction and require regional validation.
Key Points
- 1Develops random forest on 14,145 Malaysian ACS records (2006–2017), achieving AUC 0.843 overall.
- 2Highlights NOx and O₃ as leading drivers via SHAP, outperforming TIMI risk scores notably.
- 3Enables integrating pollution into clinical risk tools, potentially reclassifying NSTEMI mortality risk by 86.94%.
Scoring Rationale
Peer-reviewed regional study with large cohort and strong predictive gains, limited by single-country data and need for external validation.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
